freedreno/ir3: post-RA sched pass
After RA, we can schedule to increase parallelism (reduce nop's) without worrying about increasing register pressure. This pass lets us cut down the instruction count ~10%, and prioritize bary.f, kill, etc, which would tend to increase register pressure if we tried to do that before RA. It should be more useful if RA round-robin'd register choices. Signed-off-by:Rob Clark <robdclark@chromium.org> Part-of: <mesa/mesa!3569>
- src/freedreno/Makefile.sources 1 addition, 0 deletionssrc/freedreno/Makefile.sources
- src/freedreno/ir3/ir3.h 3 additions, 0 deletionssrc/freedreno/ir3/ir3.h
- src/freedreno/ir3/ir3_compiler_nir.c 3 additions, 0 deletionssrc/freedreno/ir3/ir3_compiler_nir.c
- src/freedreno/ir3/ir3_postsched.c 669 additions, 0 deletionssrc/freedreno/ir3/ir3_postsched.c
- src/freedreno/ir3/ir3_ra.c 2 additions, 5 deletionssrc/freedreno/ir3/ir3_ra.c
- src/freedreno/ir3/meson.build 1 addition, 0 deletionssrc/freedreno/ir3/meson.build